The origin of equine endometrial cups. II. Invasion of the endometrium by trophoblastкод для вставкиСкачать
The Origin of Equine Endometrial Cups 11. INVASION OF THE ENDOMETRIUM B Y TROPHOBLAST W. R. ALLEN, D. W. AND R. M. MOOR ARC U n i t of Reproductive Physiology a n d Biochemistry, Cambridge, England and Department of A n a t o m y , Harvard Medical School, Boston, Massachusetts 021 15 ABSTRACT Light and electron microscopic examination of tissues fixed in situ by perfusion of the gravid horn of the uteri of mares between 36 and 38 days of gestation revealed that the equine endometrial cups are composed of trophoblast cells which originate from the discrete annulate portion of the foetal membranes known as the chorionic girdle. This structure consists of closely opposed villous projections of elongated trophoblast cells and i t becomes firmly attached to the endometrium around the thirty-sixth day of pregnancy. The specialized girdle cells invade and phagocytose the endometrial epithelium and then migrate through the basal lamina into the endometrial stroma where they develop into endometrial cup cells. Measurement of pregnant mares serum gonadotrophin (PMSG) concentrations in foetal and maternal tissues of horses led Catchpole and Lyons ('34) to postulate that PMSG is secreted by the foetal chorion and stored in the endometrium. Subsequent experiments have conclusively dernonstrated, however, that PMSG is manufactured by discrete endometrial outgrowths present in the pregnant horn of the uterus of mares between the 38th and 150th days of gestation, the endometrial cups (Cole and Goss, '43; Clegg, Boda and Cole, '54). It has been widely considered in the past that the endometrial cups are entirely maternal in origin (Amoroso, '55; Gonzalez-Angullo and Hernandez-Jouregui, '71), although this hypothesis has been brought into question by the finding that foetal genotype profoundly influences PMSG levels in the maternal blood (Bielanski, Ewy and Pigoniowa, '55; Clegg, Cole, Howard and Pigon, '62; Allen, '69). Moreover, we have recently demonstrated that the only cells which possess the capacity to synthesize PMSG in vitro are those of the specialized area of the allantochorion known as the chorionic girdle (Allen and Moor, '72). The purpose of this paper is to describe the origin and histogenesis of endometrial cups at the fine-structure level. ANAT. REC., 177: 485-502. MATERIALS AND METHODS In order to maintain the delicate anatomical relationship between foetal and maternal tissues, our experiments were carried out upon material fixed in situ by perfusion of the uterus. Laparotomies were performed upon three Welsh Pony mares at 36, 37 and 38 days of gestation respectively (day of ovulation = day 0). The anaesthetized animals were positioned on their backs and a large flap of the ventral wall of the abdomen was reflected to provide direct access to the uterus and ovaries. The uterine artery to the pregnant horn was cannulated with the minimum possible handling of the uterus and perfusion was commenced. The initial pre-wash fluid of approximately 5 ml of oxygenated bicarbonate-buffered Krebs-Ringer (PH 7.4) was followed immediately by 1000 ml of 5% s-collidine buffered glutaraldehyde (pH 7.2). At first the fixative was allowed to flow unimpeded but after about 200 rnl had been perfused, the flow rate was reduced so that the remainder took about fifteen minutes to pass through the organ. Received Feb. 20, '73. Accepted June 22, '73. 1 Rockefeller Foundation Fellow in Reproductive Biology, Cambridge Universlty (1970-19711. %Presentaddress: Dr. David W. Hamilton, DeDaxtment of Anatomy, Harvard Medical School, 25 Shattuck Street, Boston, Massachusetts 02115. 485 486 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR At the end of perfusion, the uterine artery and vein were ligated and the whole uterus was carefully removed from the animal and immersed in a pan of fixative. The uterus was opened and visible endometrial cups and presumed areas of cup development were carefully dissected out with the overlying portion of allantochorion undisturbed. These pieces of tissue were then trimmed into 1 mm3 blocks and placed in fresh fixative at 4°C for one hour. New razor blades were used at each stage to minimize artifacts due to crushing. Subsequently, the blocks were exposed to 1% osmium tetroxide (pH 7.2, in s-collidine buffer), dehydrated and embedded in Epon in the usual way. Sections were cut with diamond knives and viewed with an AEI-801 electron microscope. OBSERVATIONS The clearly defined sequence of events leading to the formation of endometrial cups is shown in figure 1. Cells from a specialized area of the chorion, the chorionic girdle, attach to the endometrial epithelium and invade and destroy it by phagocytosis. The cells then migrate through the basal lamina and into the endometrial interstitium where they form definitive cups. We will describe each stage of this process in detail at the fine structural level. The chorionic girdle has been observed by others (Ewart, 1897; van Niekerk, ’65) but its significance in cup formation has not previously been suspected. A description of the development of the girdle and its fine structure is essential for an understanding of its role in endometrial cup formation. A. Development of the chorionic girdle The foetal membranes develop relatively late in gestation and close attachment with the endometrium, that is placentation, does not occur until after the 45th day of pregnancy. Up to this stage, contact between foetal and maternal tissues is confined to a small area of bilaminar omphalopleure which persists at the abembryonic pole of the conceptus. The allantois begins to develop on day 21 of gestation and by day 28 it completely surrounds the embryo and amnion. Fusion of the allantois and chorion gives rise to the allantochorion and between these layers the vascular mesoderm develops. As the allantois enlarges, the yolk sac regresses and the embryo, attached to the yolk stalk, is withdrawn towards the abembryonic pole of the conceptus (fig. 2). At the junction of the developing allantois and regressing yolk sac, the mesoderm is avascular and at this point the specialized chorionic girdle develops. On the 25th day of gestation the girdle is seen as a series of shallow corrugations of the single layer of trophoblast cells. This folding process increases so that by day 33, elongated closely apposed villous structures project from the surface of the conceptus (fig. 3). Collectively these are seen macroscopically as a discrete annulate band around the circumference of the conceptus (fig. 2). The allantois is not carried into these villous projections but rather, appears to form an anchoring point for the villi. Each villus is covered by a layer of highly modified chorion cells and the center of the villus contains clusters of these chorionic cells and other non-specific cellular elements. The modified chorion cells which comprise the epithelium of the villus (“girdle cells”) are tall columnar cells with pale nuclei. They are often binucleate and have elongated supranuclear areas. The intervillous spaces are filled with a dark staining alcian-blue-positive material which also extends over the free surface of each villus and into the lumina of adjacent endometrial glands (fig. 4). At the light microscopic level, this extracellular material often has a stippled appearance and gives the impression of an adhesive which binds foetal and maternal surfaces together. B. Fine structure of girdle cells At the electron microscopic level, the chorionic girdle cells have many features in common with mature endometrial cup cells (Hamilton, Allen and Moor, ’73). However, they also have a number of unique features which are apparently associated with their migratory function. These features disappear when the cells become fully developed, sessile, endometrial cup cells. The perinuclear regions of the highly attenuated chorionic girdle cells contain EQUINE ENDOMETRIAL CUPS. 11. 487 Fig. 1 Diagrammatic representation of the histogenesis of an endometrial cup. The chorionic girdle becomes firmly attached to the endometrium and girdle cells invade and phagocytose the endometrial epithelium. The girdle cells then migrate. through the basal lamina into the endometrial stroma where they form endometrial cup cells. 488 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR EQUINE ENDOMETRIAL CUPS. 11. primarily Golgi apparatus and associated multivesicular bodies (MVB) and vesicles (fig. 6 ) . Dictyosomes are often found completely surrounding the nucleus and, in binucleate cells, they are situated between the nuclei. The lamellae which comprise the Golgi are not fenestrated, but numerous uncoated and a few coated vesicles are usually found in the same area. The MVB have a number of curious features. In cells situated at the center of a villus (fig. 6 ) they have a very homogeneous matrix which in many respects resembles microbodies or peroxisomes from liver. They contain few vesicles. However, in the chorion cells which cover the villus and are therefore coated with extracellular material (fig. 5), the matrix of the MVB closely resembles the extracellular material itself (fig. 8). A comparison between the MVB and the extracellular material is shown in figures 7 and 8. The basal regions of girdle cells are identical to the basal regions of mature cup cells (Hamilton et al., '73). They contain profiles of short segments of endoplasmic reticulum with ribosomes scattered over its surface, scattered lipid droplets, mitochondria, and a ground substance containing free ribosomes and polyribosome rosettes. A representative section is shown in figure 9. The apical portion of the girdle cells covering the villi is highly complex (fig. 5). A few cells exhibit a relatively uniform microvillous border but in the majority, the microvilli are modified to form pseudopodia of various dimensions which extend toward the surface of the endometrium through the layer of extracellular matrix. Both the microvilli and the pseudopodia Fig. 2 Intact horse conceptus at 36 days of gestation. The arrows indicate the annulate chorionic girdle at the junction of the allantochorion and regressing yolk sac. x 1.7. Fig. 3a Chorionic girdle at 28 days of gestation. The single layer of elongated columnar girdle cells is thrown into a series of shallow ridges. x 175. Fig. 3b Chorionic girdle at 35 days of gestation. The ridges of girdle cells have now developed into elongated, branching, villous folds which project from the surface of the conceptus. Many villi have been sectioned transversely. The apical regions of the cells covering the luminal surfaces of the villi are filled with darkly staining material. x 136. 489 are filled with fine filamentous material (figs. 5, 10) which is similar in appearance to that found in certain areas along the lateral cell membrane (see fig. 6, areas of interdigitations between cells). C . Histogenesis of endometrial cups (i) Attachment of the chorionic girdle to the endometrium. At 37 days of gestation, light microscopy shows uniformly close apposition of the chorion to the endometrium. This attachment appears to be frail and is easily disturbed except in the region of the allantochorionic girdle. Here the distal extremities of the elongated chorionic cells are closely moulded to the surface of the endometrial epithelium (fig. 4) so that in areas in which some disruption has occurred through rough handling, the surfaces are mirror images of each other. (ii) Invasion of the endometrium by trophoblast. The phase of attachment of the chorionic girdle to the endometrium is apparently of short duration, for by sampling a number of potential cups from the same animal at 37 days of gestation, varying degrees of both migration and attachment are apparent. The first signs of invasion are seen as focal points of close apposition between girdle cell pseudopodia and the surface plasmalemma of endometrial epithelial cells (fig. 5). These points of apposition then broaden into more generalized areas of cell contact with resulting loss of surface microvilli from the endometrial cells and diminution of the extracellular material between the girdle cells and the endometrial cells. We have not made a detailed investigation of the nature of the junction that forms between the two cell types. The close apposition between the two plasma membranes, however, is reminiscent of the gap-type junction that has been described in many other organs (Revel and Karnovsky, '67). It is surprising to find that the chorionic cells do not invade the endometrial epithelium by way of the intercellular spaces. Rather, they appear to pass directly down the central axis of the epithelial cells. This unusual process is dramaticallv demonstrated in figure 10, where a pseudopodium from a girdle cell has intruded into the apical cytoplasm of the epithelial cell to 490 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR Fig. 4 In the light micrograph, the chorionic girdle is attached to the endometrial epithelium. The villi are obvious (arrow heads) but the extracellular space is filled with a dark-staining extracellular matrix that becomes very obvious at the mouths of the endometrial glands. x 400. EQUINE ENDOMETRIAL CUPS. 11. 49 1 Fig. 5 At the electron microscopic level one can appreciate the major fine structural features of the girdle cells. The amoeboid apical surface of the girdle cells, and the fine filamentous nature of the amoeboid processes, is contrasted sharply with the less develooed surface of endometrial epithelial cells. Note that the extracellular matrix is a t times present in apical vacuoles of the endometrial cells. X 11,790. 492 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR Fig. 6 Chorionic girdle cells are elongate and narrow, a s shown here, with complex lateral intercellular relationships. Note the texture of the cytoplasm in the intercrescent folds. x 10,900. EQUINE ENDOMETRIAL CUPS. 11. 493 Figs. 7-8 In these two micrographs, the texture of the matrix of multivesicular bodies (fig. 7) is contrasted with the extracellular matrix (fig. 8). The areas of rarefaction contain short anastomosing filaments. These areas produce a stippling effect at the light microscope level. x 34,000. 494 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR Fig. 9 The general cytoplasm of girdle cells is composed of profiles of tubular and vesicular rough endoplasmic reticulum, scattered mitochondria and numerous granules (presumably ribosomal in nature). x 14,460. EQUINE ENDOMETRIAL CUPS. 11. 495 Fig. 10 The initial stage of invasion by the girdle cells is represented here. Girdle cell processes protrude dramatically directly into the cytoplasm of the endometrial epithelial cells and do not enter the intercellular space. x 14,290. 496 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR Figs. 11-12 As girdle cells proceed through the epithelial layer they sequester the epithelial cells (fig. 1 1 ) and eventually phagocytose them (fig. 12). X 9500. EQUINE ENDOMETRIAL CUPS. 11. 497 Fig. 13 The girdle cells eventually extend their processes ( P ) through the basal lamina (arrow heads) and enter the interstitium of the endometrium. x 15,600. 498 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR Fig. 14 Wandering macrophages in the endometrial interstitium phagocytose collagen fibrils. In this illustration, collagen can be recognized in multivesicular bodies throughout the cytoplasm of the cell. x 42,000. 499 EQUINE ENDOMETRIAL CUPS. 11. the extent that the shape of the nucleus is affected. (iii) Phagocytosis of the endometrial epithelium. Intermediate stages between the processes of invasion of the chorionic girdle cells into the endometrial epithelium and their eventual migration into the interstitium are difficult to find. At the light microscopic level, endometrial epithelial cells being invaded by chorionic girdle cells are clearly visible in one area of a developing cup while in adjacent areas of the same cup, the endometrial epithelial cells have already disappeared. This is a further indication of the rapidity of the invasion process. A t the electron microscopic level, it is clear that the invading cells of the chorion sequester the epithelial cells (fig. 11) and eventually phagocytose them (fig. 12). (iv) Invasion of the endometrial stroma. The final stage in the formation of the endometrial cups involves penetration of the basal lamina of the endometrial epithelium by the invading chorionic cells and passage of the latter into the endometrial stroma. This is accomplished in the same way as the original invasion of the endometrial epithelium; pseudopodia force their way through the basal lamina, followed by the remainder of the cell. During the phagocytic phase of the invasion process, the pseudopod extensions are as numerous as in the early stages (fig. 11) and small processes from these are commonly found extending through the basal lamina (fig. 13). Larger processes are also present in the interstitium surrounded by densely packed collagen fibers (fig. 13). The endometrial stroma contains numerous fibroblasts and wandering stromal cells and just prior to girdle cell invasion, it is packed with large amounts of collagen. Relatively little collagen is found in the fully developed endometrial cup, however (Hamilton et al., '73). This decrease in collagen can probably be ascribed to the presence of numerous phagocytic cells in the interstitium which appear to be selective for collagen fibrils. One such cell, containing sequestered collagen in multivesicular-like, membrane-bounded vacuoles, is seen in figure 14. The processes of direct invasion and phagocytosis of epithelial cells described above occur only in the epithelium on the surface of the endometrium. In contrast, uterine gland epithelium is not attacked in the same manner and the invading chorionic girdle cells migrate down the length of the glands between the epithelial cells and their basal laminae without damaging or phagocytosing the epithelial cells. The girdle cells then pass through the basal lamina and into endometrial stroma where they cease to migrate, hypertrophy, and assume the epithelioid appearance of mature cup cells. DISCUSSION Although early stages of formation of the chorionic girdle have not been thoroughly studied, it is nevertheless clear that the girdle develops entirely from the chorion and contains no allantoic elements. Whether girdle development proceeds by specialization (or maturation) of preformed chorionic cells, or by production of new cells, is not clear. There are only slight structural differences between normal chorion cells that wdl go on to form the allantochorionic placenta and those that form the chorionic girdle prior to invasion. Both cells are tall, columnar in type and have amoeboid-like free surfaces. Each contains approximately the same amounts of the same cell organelles and all the chorion cells seem able to produce the alcian-blue staining extracellular material which presumably helps to bind foetal and maternal surfaces together. The one outstanding feature of the girdle cells is a large euchromatic nucleus with an immense nucleolus. This feature alone could perhaps account for the unique physiological properties of the chorionic girdle, for it implies a significant difference between nuclear-directed activities in the two populations of cells. The intracellular source of the acidic extracellular material is of some interest because of the close resemblance between this material and the matrix of the multivesicular bodies in the girdle cells. The striking similarity between the two provides only circumstantial evidence for a causal relationship between them. There is, however, some other evidence to indicate 500 W. R. ALLEN, D. W. HAMILTON AND R. M. MOOR that MVBs and lysosomes are active either in the digestion of extracellular material (Friend, '69) or in the elaboration of some of the acid hydrolases destined to be secreted by some cells (Dingle, '69; Helminnen and Ericsson, '72). It is therefore tempting to speculate that the extracellular material is secreted by the MVB of the girdle cells and that the material aids in the digestion of endometrial epithelial cell debris. Many electron micrographs show profiles of membrane debris located at the surfaces of endometrial epithelial cells in the extracellular space (fig. 5). It seems likely that these represent an initial stage in the breakdown of the epithelial plasmalemma as a result of the action of extracellular enzymes (lipases or proteinases) secreted by the girdle cells. An alternative explanation for the resemblance between MVB matrix and extracellular material is that the latter is being digested by the girdle cells. This explanation would agree with the results of studies on the uptake of exogenous tracers by other cell systems (see reviews in Dingle and Fell, '69). However there are very few coated vesicles present in chorionic girdle cells and these are considered to be essential for micropinocytosis of macromolecules (Friend and Farquhar, '67). The occurrence of alcian-blue-positive material as an extracellular matrix strongly suggests the presence of acid mucopolysaccharides. Bernfield and Banerjee ('72) and Bernfield et al. ('72) demonstrated that the extracellular acidic glycosaminoglycans play a major role in the developmental processes which determine branching patterns in mouse embryonic salivary glands. There is also evidence to indicate the importance of mucopolysaccharides in developmental processes in other systems (Manasek, '70). It is possible therefore that dependence upon extracellular mucopolysaccharides is a widespread phenomenon and that, as well as functioning as a means of attachment, the extracellular matrix may be involved in determining the developmental sequence of the invasion of the endometrium by chorionic girdle cells. By whatever means the chorionic girdle attaches to the endometrium, it is clear that active invasion of girdle cells occurs only after close apposition of the plasmalemma of the girdle cell with that of the epithelial cell. Our findings suggest that a gap junction (Revel and Karnovsky, '67) may form between these two cell types. Gap junctions are considered to be sites of low resistance contact between cells (see Goodenough and Stoeckenius, '72, for references) and they may be involved in synchronizing metabolic activity in large populations of similar cells. It is also possible that this type of junction is involved in signals between different cell types. Thus in the pregnant mare, it would be important for the foetal and maternal cells to be functionally as well as structurally interrelated, especially at the commencement of the invasion process. Such junctional specialization may therefore represent the earliest indication that implantation is about to occur. The morphological features normally associated with cell motility are abundantly represented in girdle cells. The amoeboidlike cell surface and the ectoplasmic layer of fine filaments are both features which are commonly found in actively moving cells. The feltwork nature of the f2aments seen in girdle cells (figs. 11, 12) is quite different from the discrete bundles of filaments which are described by other workers in cells such as neurulating ectoderm (Burnside, '71) and in fibroblasts (Buckley and Porter, '67). Indeed, they differ from the extensive development of filaments seen in fully developed endometrial cup cells (Hamilton et al., '73). However, the filamentous feltwork in chorionic girdle cells does closely resemble the ectoplasmic microfilaments present in growth cones of elongating axons and in migrating glial cells which Wessells et al. ('70) have shown to be involved in cell motility. It would be of interest to determine whether or not cytochalasin B, a compound that is known to affect intracellular microfilaments and cell motility (Wessells et al., '71), would prevent formation of endometrial cups in the mare. Our findings have demonstrated conclusively that endometrial cups in the mare are derived from specialized foetal trophoblast cells by a process which is summarized in figure 1. We have shown elsewhere that explants of these cells secrete gonado- EQUINE ENDOMETRIAL CUPS. 11. trophins when cultivated in vitro and when grafted allogeneically (Allen and Moor, '72). Thus, it would appear that Equidae are similar to humans and other primates in which trophoblast cells also invade the endometrium and produce gonadotrophin. There are significant differences however; in the mare, the invading trophoblast cells become completely detached from the remaining foetal membranes and they enter the endometrium some ten days before the true allantochorionic placenta is established. They remain as a discrete colony of single cells in the endometrium and do not form a syncytium. Furthermore, the invasion of lymphocytes and plasma cells which ultimately leads to the premature necrosis of the cups and their rejection from the endometrium, strongly suggests that the invading chorionic girdle cells express foetal transplantation antigens which are immLmologically recognized by the mare. ACKNOWLEDGMENTS We are grateful to Professor T. R. R. Mann and Professor D. Fawcett for their helpful criticisms of the manuscript. Financial support for this work was provided by the Thoroughbred Breeders Association and by the Rockefeller Foundation. LITERATURE CITED Allen, W. R. 1969 Factors influencing pregnant mares serum gonadotrophin production. Nature (Lond.), 223: 64-66. Allen, W. R., and R. M. Moor 1972 The origin of the equine endometrial cups. 1. Production of PMSG by fetal trophoblast cells. J. Reprod. Fert., 29: 313-316. Amoroso, E. C. 1955 Endocrinology of pregnancy. Brit. Med. Bull., 11: 117-125. Bernfield, M. R., and S . D. Banerjee 1972 Acid mucopolysaccharide (glycosaminoglycan) at the epithelial-mesenchymal interface of mouse embryo salivary glands. J. Cell Biol., 52: 664673. Bernfield, M. R., S. D. Banerjee and R. H. Cohn 1972 Dependence of salivary epithelial morphology and branching morphogenesis upon acid mucopolysaccharide-protein (proteoglycan) a t the epithelial surface. J. Cell Biol., 52: 674-689. Bielanski, W., 2. Ewy and H. Pigoniowa 1956 Difference in the level of gonadotrophin in the serum of pregnant mares. Third Intern. Congr. Anim. Reprod., Cambridge 1956, pp. 110-111. Buckley, I. K., and K. R. Porter 1967 Cytoplasmic fibrils in living cultured cells. Protoplasma, 64: 349-380. Burnside, B. 1971 Microtubules and microfila- 50 1 ments in newt neurulation. Dev. Biol., 26: 41W41. Catchpole, H. R., and W. R. Lyons 1934 The gonad-stimulating hormone of pregnant mares. Am. J. Anat., 55: 167-227. Clegg, M. T., J. M. Boda and H. H. Cole 1954 The endometrial cups and allantochorionic pouches i n the mare with emphasis on the source of equine gonadotrophin. Endocrinology, 54: 448463. Clegg, M. T., H. H. Cole, C. B. Howard and H. Pigon 1962 The influence of foetal genotype on equine gonadotrophin secretion. J. Endocr., 25: 245-248. Cole, H. H., and H. T. Goss 1943 The source of equine gonadotrophin. In: Essays in Biology i n Honour of H. M. Evans. Univ. Calif. Press, Berkeley, California, pp. 107-119. Dingle, J. T. 1969 The extracellular secretion of lysosomal enzymes. In: Lysosomes i n Biology and Medicine. J. T. Dingle and H. Fell, eds. N. Holland Publishing Co., Amsterdam, pp. 421436. Dingle, J. T., and H. Fell, eds. 1969 Lysosomes i n Biology and Medicine. N. Holland Publishing Co., Amsterdam. Ewart, J. C. 1897 A critical period in the development of the horse. Adam and Charles Black, London. Friend, D. S. 1969 Cytochemical staining of multivesicular body and Golgi vesicles. J. Cell Biol., 41: 269-279. Friend, D. S., and M. G. Farquhar 1967 Functions of coated vesicles during protein absorption i n the rat vas deferens. J. Cell. Biol., 35: 357-376. Goodenough, D. A., and W. Stoeckenius 1972 The isolation of mouse hepatocyte gap junctions. Preliminary chemical characterization and X-ray diffraction. J. Cell Biol., 54: 646-656. Gonzalez-Angulo, A., and P. Hernandez-Jaurequi 1971 The ultrastructure of secretory cells i n endometrial cups of pregnant mares. In: Twenty-ninth Ann. Meeting EMSA. pp. 506-507. Hamilton, D. W.,W. R. Allen and R. M. Moor 1973 The origin of equine endometrial cups. 111. Light and electron microscopic study of fully developed equine endometrial cups. Anat. Rec., 177: 503-518. Helminen, H. J., and J. L. E. Ericsson 1972 Ultrastructural studies o n prostatic involution in the rat. Changes in the secretory pathways. J. Ultrastr. Res., 40: 152-166. Manasek, F. J. 1970 Sulfated extracellular matrix production i n the embryonic heart and adjacent tissues. J. Exp. Zool., 174: 4 1 5 4 4 0 . Revel, J.-P., and M. J. Karnovsky 1967 Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. J. Cell Biol., 33: 7-12. Van Niekerk, C. H. 1965 The early diagnosis of pregnancy, the development of foetal membranes and nidation i n the mare. J. S. Afr. vet. med. Ass., 36: 483488. Wessells, N. K., B. S. Spooner, J. F. Ash, M. 0. Bradley, M. A. Luduana, E. L. Taylor, J. T. Wrenn and K. M. Yameda 1971 Microfilaments i n cellular and developmental process. Science, 171: 135-143.